{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "import numpy as np\n",
    "from sklearn import datasets\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import svm\n",
    "from sklearn.naive_bayes import GaussianNB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(150, 4) (150,)\n"
     ]
    }
   ],
   "source": [
    "iris = datasets.load_iris()\n",
    "X, y = iris.data, iris.target\n",
    "print(X.shape, y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def simulate(X, y, n=100):\n",
    "    X_new, y_new = [], []\n",
    "    Xmax = np.max(X, axis=0)\n",
    "    Xmin = np.min(X, axis=0)\n",
    "    for t in range(n):\n",
    "        tmp = []\n",
    "        for i in range(X.shape[1]):\n",
    "            rg = np.linspace(Xmin[i], Xmax[i], 30)\n",
    "            choice = round(np.random.choice(rg), 1)\n",
    "            tmp.append(choice)\n",
    "        X_new.append(tmp)\n",
    "        y_new.append(random.choice(list(set(y))))\n",
    "    return np.array(X_new), np.array(y_new) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_new, y_new = simulate(X, y, 1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1150, 4) (1150,)\n"
     ]
    }
   ],
   "source": [
    "X_all = np.concatenate((X, X_new), axis=0)\n",
    "y_all = np.concatenate((y, y_new), axis=0)\n",
    "print(X_all.shape, y_all.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((920, 4), (920,), (230, 4), (230,))"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_all, y_all, test_size=0.2, random_state=42)\n",
    "X_train.shape, y_train.shape, X_test.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.3869565217391304, 0.30434782608695654)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "m1 = svm.SVC(gamma='auto')\n",
    "m2 = GaussianNB()\n",
    "m1.fit(X_train, y_train);\n",
    "m2.fit(X_train, y_train);\n",
    "m1.score(X_test, y_test), m2.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def boostrap(test, m1, m2, b=10, n=10, delta=0.05):\n",
    "    ds = []\n",
    "    X_test, y_test = test\n",
    "    for i in range(b):\n",
    "        choices = random.choices(range(X_test.shape[0]), k=n)\n",
    "        _X_test = X_test[choices, :]\n",
    "        _y_test = y_test[choices]\n",
    "        dm1 = m1.score(_X_test, _y_test)\n",
    "        dm2 = m2.score(_X_test, _y_test)\n",
    "        d = dm1 - dm2\n",
    "        ds.append(d)\n",
    "    p = sum([_ for _ in ds if _ > 2*delta])/b\n",
    "    return p"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.1166666666666667"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "boostrap((X_test, y_test), m1, m2, 30, 10, 0.025)\n",
    "# 不能拒绝原假设：m1 在双边置信度 0.05 下不比 m2 好，所以 m1 不比 m2 好"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.08666666666666666"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "boostrap((X_test, y_test), m1, m2, 30, 10, 0.05)\n",
    "# 拒绝原假设：m1 在双边置信度 0.1 下不比 m2 好，所以 m1 比 m2 好"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.2"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}